Mapping facets with `vec_slice()` is too stringent

[teunbrand]

This problem was identified in #6287 and affects 3 packages.

Essentially, when you're dealing with a data.frame superclass (rather than a subclass), vec_slice() complains.

vctrs::vec_slice(mtcars, 1)
#>           mpg cyl disp  hp drat   wt  qsec vs am gear carb
#> Mazda RX4  21   6  160 110  3.9 2.62 16.46  0  1    4    4

df <- mtcars
class(df) <- union(class(df), "foo")

vctrs::vec_slice(df, 1)
#> Error in `vctrs::vec_slice()`:
#> ! `x` must be a vector, not a <data.frame/foo> object.

^{Created on 2025-01-22 with reprex v2.1.1}

This normally doesn't happen because we rebuild our own plain data.frame after evaluating aesthetics, but it becomes a problem before that when mapping facets:

ggplot2/R/facet-grid-.R

Line 312 in 97edd62

data <- vec_slice(data, facet_vals$.index)

[teunbrand]

More realistic reprex:

devtools::load_all("~/packages/ggplot2/")
#> ℹ Loading ggplot2

df <- mpg
class(df) <- union(class(df), "foo")

ggplot(df, aes(displ, hwy)) +
  geom_point() +
  facet_grid(~ year)
#> Error in `vec_slice()` at ggplot2/R/facet-grid-.R:312:7:
#> ! `x` must be a vector, not a <tbl_df/tbl/data.frame/foo> object.

^{Created on 2025-01-22 with reprex v2.1.1}